Reference
Tokdar, S. T, Xi, P., Kelly, R. and Kass, R. (2010). Journal of Computational Neuroscience, 29(1): 203-212. DOI: 10.1007/s10827-009- 0182-2.
Detection of Bursts in Neuronal Spike Trains using Hidden Semi-Markov Point Process Models. Preprint

Our MCMC algorithm adapts the forward-recursion-backward-sampling technique commonly used for hidden Markov models to the case where the state switches are semi-Markov (i.e., state durations are not memoryless). The code is written in an R-C interface.

An R package is available here.
This is built by running R CMD BUILD on this directory.

Reference
Martin, R. G. and Tokdar, S. T.(2012). Biostatistics, 13(3): 427-439.
A Nonparametric Empirical Bayes Framework for Large-scale Significance Testing. arXiv:1106.3885v3

Getting Started
C codes and a R wrapper can be found in this directory, along with many datasets. Compile the C code as
> R CMD SHLIB prml_mtest_adapt.c linpack_d.c blas1_d.c

Datasets

1. choe.rd: Golden Spike data of Choe et al (2005); see Section 6.1 of our paper.
2. golubtrain.txt Lukemia data of Golub et al (1999); see Section 6.2 of our paper
3. hedenfalk-breast-data_filt.txt: BRCA data of Hedenfalk et al (2001); see Section 6.2 of our paper

Example
Short, worked out codes are provided for all three examples in Section 6
> source("indepth_choe.R") ## will generate plots and tables for Golden spike data
> source("indepth_golub.R") ## will generate plots for Leukemia data
> source("indepth_brca.R") ## will generate plots for BRCA data

Special Note
Results may vary slightly from one run to another due to random permutation of data (usually stable). In rare cases, permutation may result in finding a local maxima of the PR marginal likelihood. A couple of runs are recommended.

Note: This was an earlier work on joint quantile regression. A more refined and comprehensive appeoach appears in Yand and Tokdar (JASA 2017) and the corresponding software may be downloaded as the R package qrjoint from CRAN.

Reference
Tokdar, S. T. and Kadane, J. B. (2012). Bayesian Analysis, 7(1): 51-72.
Simultaneous Linear Quantile Regression: A Semiparametric Bayesian Approach. Draft

C codes and a R wrapper can be found in this directory, along with some datasets.

To compile the C code use

> R CMD SHLIB slqr.c utility.c

Datasets available are:

tropcycl.rd, bwtjune97.rd.

Example
> source("slqr.r")
> load("tropcycl.rd")
> tc.slqr <- slqr(x = Year, y = WmaxST, ry = c(0, 200), nsweep = 1e4)
> ss <- 1e3 + seq(9, 9e3, 9)
> tau <- seq(.01, .99, .01)
> summary(tc.slqr, subset = ss, tau = tau, show.intercept = FALSE)

To add more sweeps to your MCMC above, you can use
> tc.slqr <- update(tc.slqr, nsweep = 1e4)

Along with summary(), other standard generics such as print(), plot(), fitted() and predict() are available.

Reference
Tokdar, S. T., Zhu, Y. M. and Ghosh, J. K. (2010). Bayesian Analysis, 5(2), 316-344.
Bayesian Density Regression with Logistic Gaussian Process and Subspace Projection. Article link -- Bayesian Analysis.

Additional Feature: Right censored response is allowed.

C codes and a R wrapper can be found in this directory, along with an R demo.

To compile the C code use

> R CMD SHLIB splgp.c utility.c